| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Date: Mon, 7 Apr 2014 15:53:47 +0800 From: Herbert Xu <herbert@...dor.apana.org.au> To: "David S. Miller" <davem@...emloft.net>, netdev@...r.kernel.org Subject: [0/2] macvlan: Handle broadcasts in work queue Hi: These two patches fix a problem where the macvlan driver tries to do too much work in netif_receive_skb. In particular, for broadcasts it'll send one skb to each macvlan interface all within a single netif_receive_skb call. This is way too much for netif_rx to handle once the number of macvlans exceeds a certain level (low thousands). Even if netif_rx could handle it, doing this amount of work in a single netif_receive_skb call seriously skews the balance of the network stack. These patches fix this by moving broadcasts to process context so that the scheduler can deal with the fairness issue. The immediate problem of the netif_rx limit is handled by using netif_rx_ni. Similar issues exist elsewhere in our stack regarding broadcasts, e.g., the bridge driver. Luckily, with the bridge driver at least we happen to have a limit of 256 ports per bridge. Still, doing 256 actions in one netif_receive_skb is less than ideal. Cheers, -- Email: Herbert Xu <herbert@...dor.apana.org.au> Home Page: http://gondor.apana.org.au/~herbert/ PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@...r.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists